US11301016B2 - Computing devices and methods of allocating power to plurality of cores in each computing device - Google Patents

Computing devices and methods of allocating power to plurality of cores in each computing device Download PDF

Info

Publication number
US11301016B2
US11301016B2 US16/518,421 US201916518421A US11301016B2 US 11301016 B2 US11301016 B2 US 11301016B2 US 201916518421 A US201916518421 A US 201916518421A US 11301016 B2 US11301016 B2 US 11301016B2
Authority
US
United States
Prior art keywords
control core
cores
processing
core
control
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active, expires
Application number
US16/518,421
Other versions
US20190339760A1 (en
Inventor
Bernhard Egger
Younghyun CHO
Su-Rim Oh
Dong-hoon Yoo
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Samsung Electronics Co Ltd
SNU R&DB Foundation
Original Assignee
Samsung Electronics Co Ltd
Seoul National University R&DB Foundation
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Samsung Electronics Co Ltd, Seoul National University R&DB Foundation filed Critical Samsung Electronics Co Ltd
Priority to US16/518,421 priority Critical patent/US11301016B2/en
Publication of US20190339760A1 publication Critical patent/US20190339760A1/en
Application granted granted Critical
Publication of US11301016B2 publication Critical patent/US11301016B2/en
Active legal-status Critical Current
Adjusted expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/28Supervision thereof, e.g. detecting power-supply failure by out of limits supervision
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/16Constructional details or arrangements
    • G06F1/20Cooling means
    • G06F1/206Cooling means comprising thermal management
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/32Means for saving power
    • G06F1/3203Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234Power saving characterised by the action undertaken
    • G06F1/3243Power saving in microcontroller unit
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5094Allocation of resources, e.g. of the central processing unit [CPU] where the allocation takes into account power or heat criteria
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/32Means for saving power
    • G06F1/3203Power management, i.e. event-based initiation of a power-saving mode
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/32Means for saving power
    • G06F1/3203Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234Power saving characterised by the action undertaken
    • G06F1/324Power saving characterised by the action undertaken by lowering clock frequency
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/32Means for saving power
    • G06F1/3203Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234Power saving characterised by the action undertaken
    • G06F1/3287Power saving characterised by the action undertaken by switching off individual functional units in the computer system
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F1/00Details not covered by groups G06F3/00 - G06F13/00 and G06F21/00
    • G06F1/26Power supply means, e.g. regulation thereof
    • G06F1/32Means for saving power
    • G06F1/3203Power management, i.e. event-based initiation of a power-saving mode
    • G06F1/3234Power saving characterised by the action undertaken
    • G06F1/3296Power saving characterised by the action undertaken by lowering the supply or operating voltage
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the present disclosure relates to computing devices each including a plurality of cores, and, to methods of dynamically allocating power to the plurality of cores.
  • the computing device is evolving to integrate a plurality of cores or processors into a single integrated circuit to meet performance requirements of applications.
  • a multi-core processor is one in which two or more cores with arithmetic functions are integrated on a single processor.
  • a many-core processor has also been developed in which more cores (usually more than 16) are integrated on a single processor.
  • the multi-core processor and the many-core processor may be mounted on portable devices such as tablet personal computers (PCs), mobile phones, personal digital assistants (PDAs), laptop computers, media players, Global Positioning System (GPS) devices, electronic book terminals, MP3 players, and digital cameras, or embedded devices including multimedia chips mounted on a televisions (TVs).
  • PCs personal computers
  • PDAs personal digital assistants
  • GPS Global Positioning System
  • MP3 players digital cameras
  • TVs televisions
  • computing devices and methods of allocating power to processing cores in a group of hierarchically classified control cores included in each computing device are provided.
  • a computing device including a plurality of cores includes a control core group including a plurality of control cores, the control core group configured to allocate a power budget to processing cores according to an energy management policy and state information of the processing cores, and transmit the allocated power budget to at least one of a lower control core and the processing cores.
  • the computing device also includes a processing core group including at least one or more of the processing cores, the processing core group configured to perform computations based on the power budget allocated by the control core group, and transmit state information of the processing cores to the control core group, the state information of the processing cores having been modified based on the computations performed; and the plurality of control cores being hierarchically classified according to a number of the processing cores to which the control cores allocate the power budget, the lower control core being a control core occupying a lower hierarchical layer than an upper control core.
  • methods of allocating power to a plurality of cores include allocating, by a control core group, a power budget to processing cores according to an energy management policy and state information of the processing cores.
  • the methods further include transmitting, by the control core group, the allocated power budget to at least one of a lower control core and the processing cores.
  • the methods further include performing, by at least one of the processing cores included in a processing core group, computations based on the allocated power budget.
  • the methods include transmitting, by the processing cores, state information of the processing cores to the control core group, the state information of the processing cores having been modified based on the computations performed; the control core group including control cores that are hierarchically classified according to a number of the processing cores to which the control cores allocate the power budget, the lower control core being a control core occupying a lower hierarchical layer than an upper control core.
  • a control core group including a plurality of control cores configured in a plurality of hierarchical levels includes a root control core in an uppermost hierarchical level, configured to obtain an energy management policy and a power budget, and transmit the energy management policy and the power budget to a plurality of leaf control cores.
  • the control core group further includes the plurality of leaf control cores in a lowermost hierarchical level, each of the plurality of leaf control cores configured to receive the energy management policy and the power budget from the root control core, and control at least one of a voltage and an operating frequency, of a subset of a plurality of processing cores, based on the received energy management policy and the received power budget.
  • FIG. 1 is a block diagram of a configuration of a computing device including a control core group, according to some example embodiments
  • FIG. 2 is a block diagram of a configuration of a computing device including a control core group that further includes a root control core and a leaf control core group, according to some example embodiments;
  • FIG. 3 illustrates the number of processing cores allocated to corresponding control cores, according to some example embodiments
  • FIG. 4 is a flow diagram illustrating the operations of a computer including control cores, according to some example embodiments.
  • FIG. 5 illustrates a structure of a computing device including control cores and a power manager, according to some example embodiments
  • FIG. 6 illustrates a structure of a processing core, according to some example embodiments.
  • FIG. 7 illustrates the interconnectivity of a leaf control core and processing cores, according to some example embodiments.
  • FIG. 8 is a flow diagram illustrating the operations of a computing device including control cores and connected to a power manager, according to some example embodiments;
  • FIG. 9 is a flowchart of methods of allocating power to a plurality of cores using a control core group, according to some example embodiments.
  • FIG. 10 is a detailed flowchart of methods of allocating power to a plurality of cores using a control core group and a power manager, according to some example embodiments.
  • FIG. 1 is a block diagram of a configuration of a computing device 100 including a control core group, according to some example embodiments.
  • the computing device 100 includes a plurality of cores, and the plurality of cores may be classified into a control core group 110 and a processing core group 120 . It is to be understood by those of ordinary skill in the art that other general-purpose components may be further included in addition to the components shown in FIG. 1 . According to some example embodiments, operations described herein as being performed by any or all of the control core group 110 and the processing core group 120 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
  • processor may refer to, for example, a hardware-implemented data processing device having circuitry that is physically structured to execute desired operations including, for example, operations represented as code and/or instructions included in a program.
  • the above-referenced hardware-implemented data processing device may include, but is not limited to, a microprocessor, a central processing unit (CPU), a processor core, a multi-core processor; a multiprocessor, an application-specific integrated circuit (ASIC), and a field programmable gate array (FPGA).
  • the control core group 110 refers to a group of control cores which allocate a power budget to a processing core, according to an energy management policy and state information of the processing core, and transmit the power budget allocated to the processing core to at least one of a lower control core (a control core, or control core group, occupying a lower hierarchical layer is hereinafter referred to as a “lower control core”) and the processing core.
  • a control core a control core, or control core group, occupying a lower hierarchical layer is hereinafter referred to as a “lower control core”
  • the description of a control core may be the same as that of the processing core. However, the description of the control core is not limited thereto as long as the control core is capable of receiving the energy management policy and the state information of the processing core and calculating the power budget allocated to the processing core.
  • each control core and processing core may be a separate processor. Operations described herein as being performed by any or all of a control core and a processing core may be performed by at least one
  • the energy management policy refers to a criterion for managing power consumption and the amount of heat generation of the computing device 100 and may be a policy based on at least one of the amount of computations of the processing core group 120 , the amount of heat generated by the processing core group 120 , and the amount of power consumed by the processing core group 120 .
  • the energy management policy may include, but is not limited to, a policy for increasing the total amount of computations of the processing core group 120 included in the computing device 100 , a policy for the processing core group 120 to perform computations based on a limited amount of power, and a policy for the processing core group 120 to perform computations based on an allowable amount of heat generation.
  • the state information of the processing core may be at least one of power consumption of the processing core, a cache miss ratio of the processing core, and information about computations performed by the processing core, but is not limited thereto as long as the state information indicates performance of the processing core.
  • the processing core group 120 refers to a group of processing cores that perform computations based on the power budget allocated by the control core group 110 and then transmit changed state information to the control core group 110 due to the computations of the processing cores.
  • a processing core may correspond to each core of the multi-core processor and the many-core processor, but is not limited thereto.
  • the plurality of control cores included in the control core group 110 may be hierarchically classified according to the number of processing cores to which the control cores allocate the power budget. An example of hierarchically classifying the control core group 110 will be described with reference to FIG. 2 .
  • FIG. 2 is a block diagram of a configuration of a computing device 200 including a control core group that further includes a root control core and a leaf control core group, according to some example embodiments.
  • the computing device 200 may include a control core group 220 and a processing core group 260 .
  • the control core group 220 may include a root control core 230 and a leaf control core group 250 .
  • the computing device 200 may further include an intermediate control core group 240 according to a hierarchical structure of the control core group 220 .
  • the processing core group 260 of FIG. 2 corresponds to the processing core group 120 of FIG. 1 , a detailed description thereof will not be given herein.
  • the root control core 230 may establish the energy management policy of the processing core group 260 and may transmit a power budget of a processing core managed by a lower control core to the lower control core.
  • FIG. 2 shows the lower control core of the root control core 230 as the intermediate control core group 240 , it is to be understood by those of ordinary skilled in the art that the lower control core of the root control core 230 may be at least one of the intermediate control core group 240 and the leaf control core group 250 , according to the hierarchical structure of the control core group 220 .
  • the root control core 230 may establish an energy management policy or receive an energy management policy from a power manager 210 based on state information of the processing core received from the lower control core.
  • the power manager 210 may be, but is not limited to, a system manager, a general-purpose Central Processing Unit (CPU), and a host processor capable of monitoring power consumption and heat generation of the computing device 200 and setting a target value for the power consumption and the heat generation of the computing device 200 .
  • the power manager 210 may be an external device to computing device 100 .
  • the leaf control core group 250 may include a plurality of leaf control cores that receive a power budget from an upper control core (a control core, or control core group, occupying a higher hierarchical layer is hereinafter referred to as an “upper control core”) and distribute power to the processing core.
  • FIG. 2 shows the upper control core of the leaf control core group 250 as the intermediate control core group 240 , it is to be understood by those of ordinary skilled in the art that the upper control core of the leaf control core group 250 may be at least one of the root control core 230 and the intermediate control core group 240 , according to the hierarchical structure of the control core group 220 .
  • the intermediate control core group 240 may include a plurality of intermediate control cores that receive a power budget from the root control core 230 and transmit a power budget of a processing core managed by the leaf control core group 250 , to the leaf control core group 250 .
  • the intermediate control core group 240 may be hierarchically classified. Referring to FIG. 2 , the intermediate control cores may be hierarchically classified according to the number of processing cores to which a power budget is allocated by the intermediate control cores.
  • an intermediate control core 241 of an upper layer may transmit a power budget of a processing core managed by an intermediate control core 242 of a lower layer, to the intermediate control core 242 of the lower layer.
  • operations described herein as being performed by any or all of the power manager 210 , the control core group 220 , the processing core group 260 , the root control core 230 , the intermediate control core group 240 , the intermediate control core 241 , the intermediate control core 242 and the leaf control core group 250 may be performed by at least one processor (e.g., the power manager 210 ) executing program code that includes instructions corresponding to the operations.
  • the instructions may be stored in a memory.
  • FIG. 3 illustrates the number of processing cores allocated to corresponding control cores, according to some example embodiments. Redundant descriptions between FIGS. 1, 2 and 3 may be omitted.
  • control cores included in the control core group 220 may be hierarchically classified according to the number of the processing cores to which the control cores allocate a power budget.
  • FIG. 3 shows a processing core group including 8 ⁇ 8 processing cores.
  • a root control core may allocate a power budget to all processing cores 310 included in the processing core group.
  • an intermediate control core may allocate a power budget to a subset 320 of the processing cores 310 included in the processing core group.
  • the number of processing cores to which one intermediate control core allocates a power budget may be 16.
  • the number of processing cores to which one intermediate control core allocates a power budget may be different for each intermediate control core.
  • a leaf control core may receive a power budget from an upper control core and control the power of the processing cores. For example, the leaf control core may allocate a power budget to a subset 330 of the processing cores 320 to which a power budget is allocated by an intermediate control core. Thus, the number of processing cores 330 to which power is distributed by the leaf control core is less than the number of the processing cores 320 to which a power budget is allocated by an intermediate control core.
  • the computing device 100 since the computing device 100 according to some example embodiments includes the control core group 110 in addition to the processing core group 120 , all the processing cores may perform computations. In addition, since the control core groups 110 are hierarchically classified, control cores may efficiently exchange information. Methods of dynamically allocating power by hierarchically classified control cores will be described with reference to FIG. 4 .
  • FIG. 4 is a flow diagram illustrating the operations of a computer 400 including control cores, according to some example embodiments.
  • the computer 400 may include a root control core 410 , an intermediate control core 420 , a leaf control core 430 and a processing core 440 .
  • operations described herein as being performed by any or all of the root control core 410 , the intermediate control core 420 , the leaf control core 430 and the processing core 440 may be performed by at least one processor executing program code that includes instructions corresponding to the operations.
  • the instructions may be stored in a memory.
  • a control core group according to some example embodiments, according to state information of a processing core received from the processing core via a lower control core, may reallocate a power budget allocated to another lower control core and another processing core.
  • a leaf control core 430 may allocate a power budget to the processing core 440 based on a power budget received from an intermediate control core 420 .
  • the leaf control core 430 may calculate the power budget allocated to the processing core 440 by taking into account state information of the processing core 440 received from the processing core 440 .
  • the leaf control core 430 may dynamically control a voltage and an operating frequency of the processing core 440 and may block power of the processing core 440 based on a power budget received from an upper control core. For example, when the leaf control core 430 receives, from an upper control core, an energy management policy that keeps the total amount of computations of the processing core 440 constant while limiting power consumption, the leaf control core 430 may block power or reduce the voltage or operating frequency of the processing core 440 .
  • the processing core 440 may collect the state information of the processing core 440 including power consumption, a cache miss ratio, the number of specific instruction words executed, and the like, and may transmit the state information to the leaf control core 430 .
  • the intermediate control core 420 may reallocate a power budget allocated to another leaf control core to the leaf control core 430 .
  • the intermediate control core 420 may also transmit the state information of the processing core 440 to the root control core 410 .
  • the root control core 410 may reallocate a power budget allocated to another intermediate control core to the processing core 440 .
  • the root control core 410 may dynamically change the energy management policy according to the state information of the processing core 440 received from the intermediate control core 420 .
  • FIG. 5 illustrates a structure of a computing device 530 including control cores and a power manager, according to some example embodiments.
  • the computing device 530 may be connected to a power manager 510 and a memory 520 . Furthermore, the computing device 530 may include a root control core 531 , a leaf control core 533 , and a processing core 534 , and may further include an intermediate control core 532 according to a hierarchical structure of a control core group. According to some example embodiments, operations described herein as being performed by any or all of the power manager 510 , the root control core 531 , the leaf control core 533 , the processing core 534 and the intermediate control core 532 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory (e.g., memory 520 ).
  • a memory e.g., memory 520
  • the root control core 531 is connected to the intermediate control core 532 . Therefore, the root control core 531 may transmit a power budget of the processing core 534 to the intermediate control core 532 instead of directly accessing the processing core 534 . Since the intermediate control core 532 is connected to the leaf control core 533 , a power budget allocated to some processing cores included in a processing core group may be transmitted to the leaf control core 533 . The leaf control core 533 may be connected to the processing core 534 to directly distribute power to the processing core 534 .
  • the root control core 531 may be connected to the power manager 510 to receive an energy management policy of the computing device 530 .
  • FIG. 6 illustrates a structure of a processing core 600 according to some example embodiments.
  • the processing core 600 may monitor its state and transmit state information to a leaf control core.
  • the processing core 600 may include a core 610 for performing computations, a router 620 for communicating with an external device, a performance monitor 630 , and a dynamic control manager 640 .
  • the performance monitor 630 may monitor operating characteristics of the core 610 and transmit monitoring results to a leaf control core via the router 620 .
  • the performance monitor 630 may monitor computations of the core 610 , the cache miss ratio of the core 610 , the number of specific computations among the computations performed by the core 610 , but information that may be monitored is not limited thereto.
  • the dynamic control manager 640 may receive information regarding a voltage and an operating frequency of the core 610 from the leaf control core via the router 620 and may control the core 610 according to the received voltage and operating frequency.
  • the processing core 600 may include a thermal sensor 650 to collect heat information.
  • the thermal sensor 650 may measure the temperature of the processing core 600 and transmit measurement results to the leaf control core via the router 620 .
  • a computing device may allocate a power budget to the processing core 600 according to an energy management policy based on the amount of heat generated by the processing core 600 .
  • operations described herein as being performed by any or all of the core 610 , the router 620 , the performance monitor 630 , the dynamic control manager 640 and the thermal sensor 650 may be performed by at least one processor executing program code that includes instructions corresponding to the operations.
  • the instructions may be stored in a memory.
  • FIG. 7 illustrates the interconnectivity of a leaf control core 730 and processing cores 710 according to some example embodiments.
  • the leaf control core 730 may distribute power to the plurality of processing cores 710 .
  • the leaf control core 730 may be connected to four processing cores 710 and may be configured to calculate a power budget of each of the processing cores 710 and to distribute power to each of the processing cores 710 based on calculation results.
  • a core 711 and a router 712 in FIG. 7 correspond to the core 610 and the router 620 in FIG. 6 , respectively, and thus a detailed description thereof will not be given herein.
  • a processing core 710 may be connected to an external router 720 via the router 712 and the external router 720 may be connected to the leaf control core 730 .
  • the leaf control core 730 may directly control the processing core 710 .
  • FIG. 7 shows that a single core 711 is included in the processing core 710 , the number of cores 711 in the processing core 710 is not limited thereto.
  • operations described herein as being performed by any or all of the processing core 710 , the core 711 , the router 712 , the external router, and the leaf control core 730 may be performed by at least one processor executing program code that includes instructions corresponding to the operations.
  • the instructions may be stored in a memory.
  • the router 712 and the external router 720 may communicate with each other according to a Network-on-Chip (NoC) communication method or a bus communication method.
  • NoC Network-on-Chip
  • a communication method of the router 712 and the external router 720 is not limited thereto.
  • FIG. 7 shows that the leaf control core 730 may distribute power to four processing cores 710 , the number of processing cores 710 to which the leaf control core 730 may distribute power is not limited thereto.
  • FIG. 8 is a flow diagram illustrating the operations of a computing device 800 including control cores and connected to a power manager according to some example embodiments.
  • the computing device 800 may be connected to a power manager 810 , a cache 820 , and an external memory 830 .
  • a root control core 840 , an intermediate control core 850 , and a leaf control core 860 of FIG. 8 correspond to the root control core 410 , the intermediate control core 420 , and the leaf control core 430 of FIG. 4 , respectively, and thus a detailed description thereof will not be given herein.
  • the root control core 840 may establish an energy management policy based on state information of a processing core 870 received from the intermediate control core 850 . In addition, the root control core 840 may receive an energy management policy from the power manager 810 .
  • the state information of the processing core 870 may include a cache miss ratio of the processing core 870 .
  • the cache miss ratio indicates a case where data or an instruction word requested by the processing core 870 is not stored in the cache 820 . If the cache miss ratio of the processing core 870 increases, the number of times the processing core 870 accesses the external memory 830 increases, and the processing core 870 consumes a higher amount of energy to perform computations.
  • operations described herein as being performed by any or all of the power manager 810 , the root control core 840 , the intermediate control core 850 , the leaf control core 860 and the processing core 870 may be performed by at least one processor (e.g., the power manager 810 ) executing program code that includes instructions corresponding to the operations.
  • the instructions may be stored in a memory (e.g., external memory 830 ).
  • FIG. 9 is a flowchart of methods of allocating power to a plurality of cores using a control core group, according to some example embodiments. Redundant descriptions between FIGS. 1 and 9 may be omitted.
  • control core group 110 may allocate a power budget to a processing core according to an energy management policy and state information of the processing core.
  • a plurality of control cores included in the control core group 110 may be hierarchically classified according to the number of processing cores to which the control cores (e.g., of control core group 110 ) allocate the power budget.
  • the energy management policy refers to a criterion for managing power consumption and the amount of heat generation of the computing device 100 .
  • the energy management policy may be a policy based on at least one of the amount of computations of the processing core group 120 , the amount of heat generated by the processing core group 120 , and the amount of power consumed by the processing core group 120 .
  • the energy management policy may include, but is not limited to, a policy for increasing the total amount of computations of the processing core group 120 included in the computing device 100 , a policy for the processing core group 120 to perform computations based on a limited amount of power, and a policy for the processing core group 120 to perform computations based on an allowable amount of heat generation.
  • the state information of the processing core may include at least one of power consumption of the processing core, a cache miss ratio of the processing core, and information about computations performed by the processing core, but is not limited thereto as long as the state information corresponds to the performance of the processing core.
  • control core group 110 may include a root control core and a leaf control core group.
  • the root control core may establish the energy management policy of the processing core group 120 and may transmit a power budget of a processing core managed by a lower control core to the lower control core. Furthermore, the root control core may establish the energy management policy based on the state information or receive the energy management policy from a power manager.
  • the leaf control core group may include at least one leaf control core that receives a power budget from the upper control core(s) and distributes power to the processing core.
  • the leaf control core may dynamically control a voltage and an operating frequency of the processing core based on a power budget received from an upper control core.
  • the leaf control core may block power of one or more processing cores included in the processing core group 120 based on the power budget received from the upper control core.
  • control core group 110 may further include an intermediate control core group.
  • An intermediate control core group is a group including a plurality of intermediate control cores that receives the power budget from the root control core and transmits the power budget of the processing core managed by the leaf control core group, to the leaf control core group.
  • the intermediate control cores are hierarchically classified according to the number of processing cores to which the power budget is allocated, the intermediate control cores of an upper layer may allocate a power budget of a processing core managed by an intermediate control core of a lower layer, to the intermediate control core of the lower layer.
  • control core group 110 may transmit the power budget allocated to the processing core to at least one of a lower control core and the processing core.
  • the processing core group 120 may perform computations based on the allocated power budget.
  • the processing core group 120 may transmit state information, changed due to the computations of the processing core, to the control core group 110 .
  • the computing device 100 may further include a thermal sensor for measuring the temperature of the processing core.
  • the state information of the processing core may further include heat generation information of the processing core received from the thermal sensor.
  • FIG. 10 is a detailed flowchart of methods of allocating power to a plurality of cores using a control core group and a power manager, according to some example embodiments.
  • FIG. 10 is a view for explaining methods of allocating power to a plurality of cores by the computing device 100 in detail, and it is to be understood by those of ordinary skilled in the art that the methods of allocating power to a plurality of cores may vary depending on a configuration of the computing device 100 .
  • operations 1020 , 1030 , 1040 , and 1050 of FIG. 10 correspond to operations 910 , 920 , 930 , and 940 of FIG. 9 , respectively, and thus a detailed description thereof will not be given herein.
  • the root control core may receive an energy management policy from the power manager.
  • the plurality of control cores (e.g., of control core group 110 ), according to state information of a processing core received from a processing core via a lower control core, may reallocate a power budget allocated to another lower control core and another processing core.
  • control cores are classified hierarchically, thereby efficiently allocating power to processing cores according to an energy management policy.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Human Computer Interaction (AREA)
  • Computer Hardware Design (AREA)
  • Computing Systems (AREA)
  • Power Sources (AREA)

Abstract

Provided are computing devices, each including a plurality of cores, and methods of allocating power to the plurality of cores. The computing device includes: a control core group including a plurality of control cores, the control core group configured to allocate a power budget to processing cores according to an energy management policy and state information of the processing cores, and transmit the allocated power budget to at least one of a lower control core and the processing cores; and a processing core group including at least one or more of the processing cores, the processing core group configured to perform computations based on the power budget allocated by the control core group, and transmit state information of the processing cores to the control core group, the state information of the processing cores having been modified based on the computations performed.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS
This application is a Continuation of U.S. application Ser. No. 15/788,293, filed on Oct. 19, 2017, which claims the benefit of Korean Patent Application No. 10-2017-025654, filed on Feb. 27, 2017, in the Korean Intellectual Property Office, the entire contents of which are incorporated herein by reference.
BACKGROUND 1. Field
The present disclosure relates to computing devices each including a plurality of cores, and, to methods of dynamically allocating power to the plurality of cores.
2. Description of the Related Art
The computing device is evolving to integrate a plurality of cores or processors into a single integrated circuit to meet performance requirements of applications. For example, a multi-core processor is one in which two or more cores with arithmetic functions are integrated on a single processor. In addition, a many-core processor has also been developed in which more cores (usually more than 16) are integrated on a single processor. The multi-core processor and the many-core processor may be mounted on portable devices such as tablet personal computers (PCs), mobile phones, personal digital assistants (PDAs), laptop computers, media players, Global Positioning System (GPS) devices, electronic book terminals, MP3 players, and digital cameras, or embedded devices including multimedia chips mounted on a televisions (TVs).
It may be desirable to devise methods of managing power and heat-generation of cores in a multi-core processor or many-core processor.
SUMMARY
Provided are computing devices and methods of allocating power to processing cores in a group of hierarchically classified control cores included in each computing device.
Additional aspects will be set forth in part in the description which follows and, in part, will be apparent from the description, or may be learned by practice of the presented example embodiments.
According to some example embodiments, a computing device including a plurality of cores includes a control core group including a plurality of control cores, the control core group configured to allocate a power budget to processing cores according to an energy management policy and state information of the processing cores, and transmit the allocated power budget to at least one of a lower control core and the processing cores. The computing device also includes a processing core group including at least one or more of the processing cores, the processing core group configured to perform computations based on the power budget allocated by the control core group, and transmit state information of the processing cores to the control core group, the state information of the processing cores having been modified based on the computations performed; and the plurality of control cores being hierarchically classified according to a number of the processing cores to which the control cores allocate the power budget, the lower control core being a control core occupying a lower hierarchical layer than an upper control core.
According to some example embodiments, methods of allocating power to a plurality of cores include allocating, by a control core group, a power budget to processing cores according to an energy management policy and state information of the processing cores. The methods further include transmitting, by the control core group, the allocated power budget to at least one of a lower control core and the processing cores. The methods further include performing, by at least one of the processing cores included in a processing core group, computations based on the allocated power budget. Furthermore, the methods include transmitting, by the processing cores, state information of the processing cores to the control core group, the state information of the processing cores having been modified based on the computations performed; the control core group including control cores that are hierarchically classified according to a number of the processing cores to which the control cores allocate the power budget, the lower control core being a control core occupying a lower hierarchical layer than an upper control core.
According to some example embodiments, a control core group including a plurality of control cores configured in a plurality of hierarchical levels includes a root control core in an uppermost hierarchical level, configured to obtain an energy management policy and a power budget, and transmit the energy management policy and the power budget to a plurality of leaf control cores. The control core group further includes the plurality of leaf control cores in a lowermost hierarchical level, each of the plurality of leaf control cores configured to receive the energy management policy and the power budget from the root control core, and control at least one of a voltage and an operating frequency, of a subset of a plurality of processing cores, based on the received energy management policy and the received power budget.
BRIEF DESCRIPTION OF THE DRAWINGS
These and/or other aspects will become apparent and more readily appreciated from the following description of some example embodiments, taken in conjunction with the accompanying drawings in which:
FIG. 1 is a block diagram of a configuration of a computing device including a control core group, according to some example embodiments;
FIG. 2 is a block diagram of a configuration of a computing device including a control core group that further includes a root control core and a leaf control core group, according to some example embodiments;
FIG. 3 illustrates the number of processing cores allocated to corresponding control cores, according to some example embodiments;
FIG. 4 is a flow diagram illustrating the operations of a computer including control cores, according to some example embodiments;
FIG. 5 illustrates a structure of a computing device including control cores and a power manager, according to some example embodiments;
FIG. 6 illustrates a structure of a processing core, according to some example embodiments;
FIG. 7 illustrates the interconnectivity of a leaf control core and processing cores, according to some example embodiments;
FIG. 8 is a flow diagram illustrating the operations of a computing device including control cores and connected to a power manager, according to some example embodiments;
FIG. 9 is a flowchart of methods of allocating power to a plurality of cores using a control core group, according to some example embodiments; and
FIG. 10 is a detailed flowchart of methods of allocating power to a plurality of cores using a control core group and a power manager, according to some example embodiments.
DETAILED DESCRIPTION
Reference will now be made in detail to some example embodiments, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like elements throughout. In this regard, the present example embodiments may have different forms and should not be construed as being limited to the descriptions set forth herein. Accordingly, the example embodiments are merely described below, by referring to the figures, to explain aspects. Expressions such as “at least one of,” when preceding a list of elements, modify the entire list of elements and do not modify the individual elements of the list.
FIG. 1 is a block diagram of a configuration of a computing device 100 including a control core group, according to some example embodiments.
The computing device 100 according to some example embodiments includes a plurality of cores, and the plurality of cores may be classified into a control core group 110 and a processing core group 120. It is to be understood by those of ordinary skill in the art that other general-purpose components may be further included in addition to the components shown in FIG. 1. According to some example embodiments, operations described herein as being performed by any or all of the control core group 110 and the processing core group 120 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory. The term ‘processor,’ as used in the present disclosure, may refer to, for example, a hardware-implemented data processing device having circuitry that is physically structured to execute desired operations including, for example, operations represented as code and/or instructions included in a program. In at least some example embodiments the above-referenced hardware-implemented data processing device may include, but is not limited to, a microprocessor, a central processing unit (CPU), a processor core, a multi-core processor; a multiprocessor, an application-specific integrated circuit (ASIC), and a field programmable gate array (FPGA).
The control core group 110 refers to a group of control cores which allocate a power budget to a processing core, according to an energy management policy and state information of the processing core, and transmit the power budget allocated to the processing core to at least one of a lower control core (a control core, or control core group, occupying a lower hierarchical layer is hereinafter referred to as a “lower control core”) and the processing core. Here, the description of a control core may be the same as that of the processing core. However, the description of the control core is not limited thereto as long as the control core is capable of receiving the energy management policy and the state information of the processing core and calculating the power budget allocated to the processing core. According to some example embodiments, each control core and processing core may be a separate processor. Operations described herein as being performed by any or all of a control core and a processing core may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
Meanwhile, the energy management policy refers to a criterion for managing power consumption and the amount of heat generation of the computing device 100 and may be a policy based on at least one of the amount of computations of the processing core group 120, the amount of heat generated by the processing core group 120, and the amount of power consumed by the processing core group 120. In more detail, the energy management policy may include, but is not limited to, a policy for increasing the total amount of computations of the processing core group 120 included in the computing device 100, a policy for the processing core group 120 to perform computations based on a limited amount of power, and a policy for the processing core group 120 to perform computations based on an allowable amount of heat generation. Also, the state information of the processing core may be at least one of power consumption of the processing core, a cache miss ratio of the processing core, and information about computations performed by the processing core, but is not limited thereto as long as the state information indicates performance of the processing core.
The processing core group 120 refers to a group of processing cores that perform computations based on the power budget allocated by the control core group 110 and then transmit changed state information to the control core group 110 due to the computations of the processing cores. Here, a processing core may correspond to each core of the multi-core processor and the many-core processor, but is not limited thereto.
The plurality of control cores included in the control core group 110 may be hierarchically classified according to the number of processing cores to which the control cores allocate the power budget. An example of hierarchically classifying the control core group 110 will be described with reference to FIG. 2.
FIG. 2 is a block diagram of a configuration of a computing device 200 including a control core group that further includes a root control core and a leaf control core group, according to some example embodiments.
Referring to FIG. 2, the computing device 200 may include a control core group 220 and a processing core group 260. The control core group 220 may include a root control core 230 and a leaf control core group 250. In addition, the computing device 200 may further include an intermediate control core group 240 according to a hierarchical structure of the control core group 220. Meanwhile, since the processing core group 260 of FIG. 2 corresponds to the processing core group 120 of FIG. 1, a detailed description thereof will not be given herein.
The root control core 230 may establish the energy management policy of the processing core group 260 and may transmit a power budget of a processing core managed by a lower control core to the lower control core. Although FIG. 2 shows the lower control core of the root control core 230 as the intermediate control core group 240, it is to be understood by those of ordinary skilled in the art that the lower control core of the root control core 230 may be at least one of the intermediate control core group 240 and the leaf control core group 250, according to the hierarchical structure of the control core group 220.
Meanwhile, the root control core 230 may establish an energy management policy or receive an energy management policy from a power manager 210 based on state information of the processing core received from the lower control core. The power manager 210 may be, but is not limited to, a system manager, a general-purpose Central Processing Unit (CPU), and a host processor capable of monitoring power consumption and heat generation of the computing device 200 and setting a target value for the power consumption and the heat generation of the computing device 200. In some example embodiments, the power manager 210 may be an external device to computing device 100.
The leaf control core group 250 may include a plurality of leaf control cores that receive a power budget from an upper control core (a control core, or control core group, occupying a higher hierarchical layer is hereinafter referred to as an “upper control core”) and distribute power to the processing core. Although FIG. 2 shows the upper control core of the leaf control core group 250 as the intermediate control core group 240, it is to be understood by those of ordinary skilled in the art that the upper control core of the leaf control core group 250 may be at least one of the root control core 230 and the intermediate control core group 240, according to the hierarchical structure of the control core group 220.
The intermediate control core group 240 may include a plurality of intermediate control cores that receive a power budget from the root control core 230 and transmit a power budget of a processing core managed by the leaf control core group 250, to the leaf control core group 250. The intermediate control core group 240 may be hierarchically classified. Referring to FIG. 2, the intermediate control cores may be hierarchically classified according to the number of processing cores to which a power budget is allocated by the intermediate control cores. Here, an intermediate control core 241 of an upper layer may transmit a power budget of a processing core managed by an intermediate control core 242 of a lower layer, to the intermediate control core 242 of the lower layer. According to some example embodiments, operations described herein as being performed by any or all of the power manager 210, the control core group 220, the processing core group 260, the root control core 230, the intermediate control core group 240, the intermediate control core 241, the intermediate control core 242 and the leaf control core group 250 may be performed by at least one processor (e.g., the power manager 210) executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
FIG. 3 illustrates the number of processing cores allocated to corresponding control cores, according to some example embodiments. Redundant descriptions between FIGS. 1, 2 and 3 may be omitted.
According to some example embodiments, control cores included in the control core group 220 may be hierarchically classified according to the number of the processing cores to which the control cores allocate a power budget.
FIG. 3 shows a processing core group including 8×8 processing cores. Referring to FIG. 3, a root control core may allocate a power budget to all processing cores 310 included in the processing core group.
Meanwhile, an intermediate control core may allocate a power budget to a subset 320 of the processing cores 310 included in the processing core group. For example, if the intermediate control core group 240 includes four intermediate control cores in a single layer, the number of processing cores to which one intermediate control core allocates a power budget may be 16. However, it is to be understood by those of ordinary skilled in the art that the number of processing cores to which one intermediate control core allocates a power budget may be different for each intermediate control core.
Meanwhile, a leaf control core may receive a power budget from an upper control core and control the power of the processing cores. For example, the leaf control core may allocate a power budget to a subset 330 of the processing cores 320 to which a power budget is allocated by an intermediate control core. Thus, the number of processing cores 330 to which power is distributed by the leaf control core is less than the number of the processing cores 320 to which a power budget is allocated by an intermediate control core.
In addition, since the computing device 100 according to some example embodiments includes the control core group 110 in addition to the processing core group 120, all the processing cores may perform computations. In addition, since the control core groups 110 are hierarchically classified, control cores may efficiently exchange information. Methods of dynamically allocating power by hierarchically classified control cores will be described with reference to FIG. 4.
FIG. 4 is a flow diagram illustrating the operations of a computer 400 including control cores, according to some example embodiments. The computer 400 may include a root control core 410, an intermediate control core 420, a leaf control core 430 and a processing core 440. According to some example embodiments, operations described herein as being performed by any or all of the root control core 410, the intermediate control core 420, the leaf control core 430 and the processing core 440 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
A control core group according to some example embodiments, according to state information of a processing core received from the processing core via a lower control core, may reallocate a power budget allocated to another lower control core and another processing core.
Referring to FIG. 4, a leaf control core 430 may allocate a power budget to the processing core 440 based on a power budget received from an intermediate control core 420. The leaf control core 430 may calculate the power budget allocated to the processing core 440 by taking into account state information of the processing core 440 received from the processing core 440.
Meanwhile, the leaf control core 430 may dynamically control a voltage and an operating frequency of the processing core 440 and may block power of the processing core 440 based on a power budget received from an upper control core. For example, when the leaf control core 430 receives, from an upper control core, an energy management policy that keeps the total amount of computations of the processing core 440 constant while limiting power consumption, the leaf control core 430 may block power or reduce the voltage or operating frequency of the processing core 440.
Meanwhile, the processing core 440 may collect the state information of the processing core 440 including power consumption, a cache miss ratio, the number of specific instruction words executed, and the like, and may transmit the state information to the leaf control core 430.
The intermediate control core 420, according to an energy management policy and a power budget received from a root control core 410 and the state information of the processing core 440 received from the leaf control core 430, may reallocate a power budget allocated to another leaf control core to the leaf control core 430. The intermediate control core 420 may also transmit the state information of the processing core 440 to the root control core 410.
The root control core 410, according to the state information of the processing core 440 received from the intermediate control core 420, may reallocate a power budget allocated to another intermediate control core to the processing core 440. In addition, the root control core 410 may dynamically change the energy management policy according to the state information of the processing core 440 received from the intermediate control core 420.
FIG. 5 illustrates a structure of a computing device 530 including control cores and a power manager, according to some example embodiments.
Referring to FIG. 5, the computing device 530 may be connected to a power manager 510 and a memory 520. Furthermore, the computing device 530 may include a root control core 531, a leaf control core 533, and a processing core 534, and may further include an intermediate control core 532 according to a hierarchical structure of a control core group. According to some example embodiments, operations described herein as being performed by any or all of the power manager 510, the root control core 531, the leaf control core 533, the processing core 534 and the intermediate control core 532 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory (e.g., memory 520).
Referring to FIG. 5, the root control core 531 is connected to the intermediate control core 532. Therefore, the root control core 531 may transmit a power budget of the processing core 534 to the intermediate control core 532 instead of directly accessing the processing core 534. Since the intermediate control core 532 is connected to the leaf control core 533, a power budget allocated to some processing cores included in a processing core group may be transmitted to the leaf control core 533. The leaf control core 533 may be connected to the processing core 534 to directly distribute power to the processing core 534.
Meanwhile, the root control core 531 may be connected to the power manager 510 to receive an energy management policy of the computing device 530.
FIG. 6 illustrates a structure of a processing core 600 according to some example embodiments.
The processing core 600 according to some example embodiments may monitor its state and transmit state information to a leaf control core. For example, referring to FIG. 6, the processing core 600 may include a core 610 for performing computations, a router 620 for communicating with an external device, a performance monitor 630, and a dynamic control manager 640.
The performance monitor 630 may monitor operating characteristics of the core 610 and transmit monitoring results to a leaf control core via the router 620. For example, the performance monitor 630 may monitor computations of the core 610, the cache miss ratio of the core 610, the number of specific computations among the computations performed by the core 610, but information that may be monitored is not limited thereto.
The dynamic control manager 640 may receive information regarding a voltage and an operating frequency of the core 610 from the leaf control core via the router 620 and may control the core 610 according to the received voltage and operating frequency.
In addition, the processing core 600 according to some example embodiments may include a thermal sensor 650 to collect heat information. The thermal sensor 650 may measure the temperature of the processing core 600 and transmit measurement results to the leaf control core via the router 620. For example, if the processing core 600 includes the thermal sensor 650, a computing device may allocate a power budget to the processing core 600 according to an energy management policy based on the amount of heat generated by the processing core 600. According to some example embodiments, operations described herein as being performed by any or all of the core 610, the router 620, the performance monitor 630, the dynamic control manager 640 and the thermal sensor 650 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
FIG. 7 illustrates the interconnectivity of a leaf control core 730 and processing cores 710 according to some example embodiments.
The leaf control core 730 according to some example embodiments may distribute power to the plurality of processing cores 710. For example, referring to FIG. 7, the leaf control core 730 may be connected to four processing cores 710 and may be configured to calculate a power budget of each of the processing cores 710 and to distribute power to each of the processing cores 710 based on calculation results. A core 711 and a router 712 in FIG. 7 correspond to the core 610 and the router 620 in FIG. 6, respectively, and thus a detailed description thereof will not be given herein.
A processing core 710 may be connected to an external router 720 via the router 712 and the external router 720 may be connected to the leaf control core 730. Thus, the leaf control core 730 may directly control the processing core 710. Although FIG. 7 shows that a single core 711 is included in the processing core 710, the number of cores 711 in the processing core 710 is not limited thereto. According to some example embodiments, operations described herein as being performed by any or all of the processing core 710, the core 711, the router 712, the external router, and the leaf control core 730 may be performed by at least one processor executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory.
The router 712 and the external router 720 may communicate with each other according to a Network-on-Chip (NoC) communication method or a bus communication method. However, a communication method of the router 712 and the external router 720 is not limited thereto.
Although FIG. 7 shows that the leaf control core 730 may distribute power to four processing cores 710, the number of processing cores 710 to which the leaf control core 730 may distribute power is not limited thereto.
FIG. 8 is a flow diagram illustrating the operations of a computing device 800 including control cores and connected to a power manager according to some example embodiments.
The computing device 800 according to some example embodiments may be connected to a power manager 810, a cache 820, and an external memory 830. Meanwhile, a root control core 840, an intermediate control core 850, and a leaf control core 860 of FIG. 8 correspond to the root control core 410, the intermediate control core 420, and the leaf control core 430 of FIG. 4, respectively, and thus a detailed description thereof will not be given herein.
The root control core 840 may establish an energy management policy based on state information of a processing core 870 received from the intermediate control core 850. In addition, the root control core 840 may receive an energy management policy from the power manager 810.
Meanwhile, the state information of the processing core 870 may include a cache miss ratio of the processing core 870. The cache miss ratio indicates a case where data or an instruction word requested by the processing core 870 is not stored in the cache 820. If the cache miss ratio of the processing core 870 increases, the number of times the processing core 870 accesses the external memory 830 increases, and the processing core 870 consumes a higher amount of energy to perform computations. According to some example embodiments, operations described herein as being performed by any or all of the power manager 810, the root control core 840, the intermediate control core 850, the leaf control core 860 and the processing core 870 may be performed by at least one processor (e.g., the power manager 810) executing program code that includes instructions corresponding to the operations. The instructions may be stored in a memory (e.g., external memory 830).
FIG. 9 is a flowchart of methods of allocating power to a plurality of cores using a control core group, according to some example embodiments. Redundant descriptions between FIGS. 1 and 9 may be omitted.
In operation 910, the control core group 110 may allocate a power budget to a processing core according to an energy management policy and state information of the processing core. A plurality of control cores included in the control core group 110 may be hierarchically classified according to the number of processing cores to which the control cores (e.g., of control core group 110) allocate the power budget.
The energy management policy refers to a criterion for managing power consumption and the amount of heat generation of the computing device 100. The energy management policy may be a policy based on at least one of the amount of computations of the processing core group 120, the amount of heat generated by the processing core group 120, and the amount of power consumed by the processing core group 120. In more detail, the energy management policy may include, but is not limited to, a policy for increasing the total amount of computations of the processing core group 120 included in the computing device 100, a policy for the processing core group 120 to perform computations based on a limited amount of power, and a policy for the processing core group 120 to perform computations based on an allowable amount of heat generation.
Also, the state information of the processing core may include at least one of power consumption of the processing core, a cache miss ratio of the processing core, and information about computations performed by the processing core, but is not limited thereto as long as the state information corresponds to the performance of the processing core.
In more detail, the control core group 110 may include a root control core and a leaf control core group.
The root control core may establish the energy management policy of the processing core group 120 and may transmit a power budget of a processing core managed by a lower control core to the lower control core. Furthermore, the root control core may establish the energy management policy based on the state information or receive the energy management policy from a power manager.
The leaf control core group may include at least one leaf control core that receives a power budget from the upper control core(s) and distributes power to the processing core. In more detail, the leaf control core may dynamically control a voltage and an operating frequency of the processing core based on a power budget received from an upper control core. In addition, the leaf control core may block power of one or more processing cores included in the processing core group 120 based on the power budget received from the upper control core.
In addition, the control core group 110 may further include an intermediate control core group. An intermediate control core group is a group including a plurality of intermediate control cores that receives the power budget from the root control core and transmits the power budget of the processing core managed by the leaf control core group, to the leaf control core group. When the intermediate control cores are hierarchically classified according to the number of processing cores to which the power budget is allocated, the intermediate control cores of an upper layer may allocate a power budget of a processing core managed by an intermediate control core of a lower layer, to the intermediate control core of the lower layer.
In operation 920, the control core group 110 may transmit the power budget allocated to the processing core to at least one of a lower control core and the processing core.
In operation 930, the processing core group 120, including at least one of the processing cores, may perform computations based on the allocated power budget.
In operation 940, the processing core group 120 may transmit state information, changed due to the computations of the processing core, to the control core group 110.
Meanwhile, the computing device 100 may further include a thermal sensor for measuring the temperature of the processing core. As a result, the state information of the processing core may further include heat generation information of the processing core received from the thermal sensor.
FIG. 10 is a detailed flowchart of methods of allocating power to a plurality of cores using a control core group and a power manager, according to some example embodiments.
FIG. 10 is a view for explaining methods of allocating power to a plurality of cores by the computing device 100 in detail, and it is to be understood by those of ordinary skilled in the art that the methods of allocating power to a plurality of cores may vary depending on a configuration of the computing device 100. Furthermore, operations 1020, 1030, 1040, and 1050 of FIG. 10 correspond to operations 910, 920, 930, and 940 of FIG. 9, respectively, and thus a detailed description thereof will not be given herein.
In operation 1010, the root control core may receive an energy management policy from the power manager.
In operation 1060, the plurality of control cores (e.g., of control core group 110), according to state information of a processing core received from a processing core via a lower control core, may reallocate a power budget allocated to another lower control core and another processing core.
According to some example embodiments, control cores are classified hierarchically, thereby efficiently allocating power to processing cores according to an energy management policy.
It should be understood that example embodiments described herein should be considered in a descriptive sense only and not for purposes of limitation. Descriptions of features or aspects within example embodiments should typically be considered as available for other similar features or aspects in other example embodiments.
While one or more example embodiments have been described with reference to the figures, it will be understood by those of ordinary skill in the art that various changes in form and details may be made therein without departing from the spirit and scope as defined by the following claims.

Claims (20)

What is claimed is:
1. A computing device including a plurality of cores, the computing device comprising:
a control core group including a plurality of control cores, the control core group including a root control core configured to
obtain an energy management policy from an external source,
allocate a power budget to processing cores according to the energy management policy and state information of the processing cores, and
transmit the allocated power budget to a lower control core of the plurality of control cores; and
a processing core group including at least one or more of the processing cores, the processing core group configured to perform computations based on the power budget; and
the plurality of control cores being hierarchically classified according to a number of the processing cores to which the control cores allocate the power budget, the lower control core occupying a lower hierarchical layer than an upper control core of the plurality of control cores.
2. The computing device of claim 1, wherein the control core group further includes:
a leaf control core group including a plurality of leaf control cores, the leaf control core group configured to
receive the power budget from an upper control core and
distribute power to the processing cores.
3. The computing device of claim 2, wherein
the control core group is further configured to
reallocate a power budget allocated to another lower control core and other processing cores according to state information of the processing cores received from the lower control core and the processing cores.
4. The computing device of claim 3, wherein
the state information is at least one of power consumption of the processing cores, cache miss ratios of the processing cores, and computations performed by the processing cores.
5. The computing device of claim 2, wherein the control core group further includes:
an intermediate control core group, including a plurality of intermediate control cores, the intermediate control core group configured to
receive the power budget from the root control core, and
transmit the power budget of the processing cores managed by the leaf control core group to the leaf control core group;
the intermediate control cores being hierarchically classified according to the number of the processing cores to which the intermediate control cores allocate the power budget; and
an intermediate control core of an upper hierarchical layer transmitting the power budget of the processing cores managed by an intermediate control core of a lower hierarchical layer, to the intermediate control core of the lower hierarchical layer.
6. The computing device of claim 2, wherein
the root control core is further configured to receive the energy management policy from a power manager.
7. The computing device of claim 2, wherein
the leaf control core group is further configured to
dynamically control voltages and operating frequencies of the processing cores based on the power budget received from the upper control core.
8. The computing device of claim 2, wherein
the leaf control core group is further configured to
block power of one or more of the processing cores included in the processing core group based on the power budget received from the upper control core.
9. The computing device of claim 1, further comprising:
a thermal sensor configured to measure temperatures of the processing cores; and
the state information further including heat generation information of the processing cores received from the thermal sensor.
10. The computing device of claim 9, wherein
the energy management policy is based on at least one of computations of the processing core group, an amount of heat generated by the processing core group, and an amount of power used by the processing core group.
11. A computing device including a plurality of cores, the computing device comprising:
at least one root control core configured to
obtain an energy management policy from an external source,
receive state information of at least one processing core,
allocate a power budget to the at least one processing core according to the energy management policy and the state information, and
transmit the allocated power budget to at least one intermediate control core;
at least one leaf control core configured to receive the power budget from the at least one intermediate control core and distribute power to at least one processing core; and
the at least one processing core configured to perform computations based on the power distributed by the at least one leaf control core.
12. The computing device of claim 11, wherein
the at least one intermediate control core includes a plurality of hierarchical layers, and
an intermediate control core of an uppermost layer transmits the power budget to an intermediate control core of a lowest layer.
13. The computing device of claim 11, wherein
a number of first processing cores to which power is distributed by the at least one leaf control core is less than a number of second processing cores to which a power budget is distributed by the at least one intermediate control core.
14. The computing device of claim 11, wherein
the at least one root control core, the at least one intermediate control core and the at least one leaf control core are hierarchically classified based on a number of the at least one processing cores.
15. The computing device of claim 11, wherein
the state information of the at least one processing core includes at least one of power consumption of the processing core, a cache miss ratio of the processing core and information of computations performed by the processing core.
16. A computing device including a plurality of cores, the computing device comprising:
a control core group including a plurality of control cores, the control core group including a root control core configured to
obtain an energy management policy from an external source,
allocate a first power budget to processing cores according to the energy management policy and first state information of the processing cores, and
transmit the first power budget to a lower control core of
the plurality of control cores; and
a processing core group including at least one or more of the processing cores, the processing core group configured to
perform computations based on the first power budget allocated by the control core group, and
transmit second state information of the processing cores to the control core group, the second state information having been modified from the first state information based on the computations performed; and
the control core group being configured to allocate a second power budget to processing cores according to the second state information, the plurality of control cores being hierarchically classified according to a number of the processing cores.
17. The computing device of claim 16, wherein the control core group further includes:
a leaf control core group including a plurality of leaf control cores, the leaf control core group configured to
receive the first power budget from an upper control core and
distribute power to the processing cores.
18. The computing device of claim 17, wherein
the processing core group is configured to transmit the second state information to the leaf control core group; and
the leaf control core group is configured to transmit the second state information to the root control core.
19. The computing device of claim 17, further comprising:
a intermediate core group including a plurality of intermediate cores, communicates between the root control core and the leaf control core group.
20. The computing device of claim 16, further comprising:
a thermal sensor configured to measure temperatures of the processing cores,
wherein the first state information includes heat generation information of the processing cores received from the thermal sensor.
US16/518,421 2017-02-27 2019-07-22 Computing devices and methods of allocating power to plurality of cores in each computing device Active 2038-05-07 US11301016B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/518,421 US11301016B2 (en) 2017-02-27 2019-07-22 Computing devices and methods of allocating power to plurality of cores in each computing device

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
KR10-2017-0025654 2017-02-27
KR1020170025654A KR20180098904A (en) 2017-02-27 2017-02-27 Computing device and method for allocating power to the plurality of cores in the computing device
US15/788,293 US10409351B2 (en) 2017-02-27 2017-10-19 Computing devices and methods of allocating power to plurality of cores in each computing device
US16/518,421 US11301016B2 (en) 2017-02-27 2019-07-22 Computing devices and methods of allocating power to plurality of cores in each computing device

Related Parent Applications (1)

Application Number Title Priority Date Filing Date
US15/788,293 Continuation US10409351B2 (en) 2017-02-27 2017-10-19 Computing devices and methods of allocating power to plurality of cores in each computing device

Publications (2)

Publication Number Publication Date
US20190339760A1 US20190339760A1 (en) 2019-11-07
US11301016B2 true US11301016B2 (en) 2022-04-12

Family

ID=63246768

Family Applications (2)

Application Number Title Priority Date Filing Date
US15/788,293 Active 2037-12-25 US10409351B2 (en) 2017-02-27 2017-10-19 Computing devices and methods of allocating power to plurality of cores in each computing device
US16/518,421 Active 2038-05-07 US11301016B2 (en) 2017-02-27 2019-07-22 Computing devices and methods of allocating power to plurality of cores in each computing device

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US15/788,293 Active 2037-12-25 US10409351B2 (en) 2017-02-27 2017-10-19 Computing devices and methods of allocating power to plurality of cores in each computing device

Country Status (3)

Country Link
US (2) US10409351B2 (en)
KR (1) KR20180098904A (en)
CN (1) CN108509014B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11174329B2 (en) * 2016-09-30 2021-11-16 Dow Global Technologies Llc Multi- or dual-headed compositions useful for chain shuttling and process to prepare the same
US11567556B2 (en) * 2019-03-28 2023-01-31 Intel Corporation Platform slicing of central processing unit (CPU) resources
US11698812B2 (en) * 2019-08-29 2023-07-11 Intel Corporation System, apparatus and method for providing hardware state feedback to an operating system in a heterogeneous processor
CN113360186A (en) * 2020-03-04 2021-09-07 北京希姆计算科技有限公司 Task scheduling method and device, electronic equipment and computer readable storage medium
CN115048194A (en) * 2021-03-09 2022-09-13 华为技术有限公司 Task allocation method and device
US20220318056A1 (en) * 2021-03-30 2022-10-06 Advanced Micro Devices, Inc. Dynamic system power load management
US11797045B2 (en) 2021-09-22 2023-10-24 Qualcomm Incorporated Dynamic voltage and frequency scaling (DVFS) within processor clusters
US11662795B2 (en) * 2021-09-29 2023-05-30 Dell Products L.P. System and method for ending a power reduction protocol in an information handling system environment

Citations (38)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6372608B1 (en) 1996-08-27 2002-04-16 Seiko Epson Corporation Separating method, method for transferring thin film device, thin film device, thin film integrated circuit device, and liquid crystal display device manufactured by using the transferring method
USRE38466E1 (en) 1996-11-12 2004-03-16 Seiko Epson Corporation Manufacturing method of active matrix substrate, active matrix substrate and liquid crystal display device
US6711691B1 (en) 1999-05-13 2004-03-23 Apple Computer, Inc. Power management for computer systems
US6818465B2 (en) 2001-08-22 2004-11-16 Sony Corporation Nitride semiconductor element and production method for nitride semiconductor element
US6858081B2 (en) 2002-01-17 2005-02-22 Sony Corporation Selective growth method, and semiconductor light emitting device and fabrication method thereof
US6967353B2 (en) 2002-01-18 2005-11-22 Sony Corporation Semiconductor light emitting device and fabrication method thereof
US7002182B2 (en) 2002-09-06 2006-02-21 Sony Corporation Semiconductor light emitting device integral type semiconductor light emitting unit image display unit and illuminating unit
US7084420B2 (en) 2004-10-26 2006-08-01 Samsung Electro-Mechanics Co., Ltd. Nitride based semiconductor device
US7087932B2 (en) 2000-07-18 2006-08-08 Sony Corporation Semiconductor light-emitting device and semiconductor light-emitting device
US7154124B2 (en) 2003-12-23 2006-12-26 Samsung Electro-Mechanics Co., Ltd. Nitride semiconductor light emitting device and method for manufacturing the same
US7208725B2 (en) 1998-11-25 2007-04-24 Rohm And Haas Electronic Materials Llc Optoelectronic component with encapsulant
US20070198863A1 (en) * 2006-02-17 2007-08-23 Pradip Bose Method and system for controlling power in a chip through a power-performance monitor and control unit
US7281146B2 (en) 2004-06-30 2007-10-09 Intel Corporation Dynamic power requirement budget manager
US7501656B2 (en) 2005-07-26 2009-03-10 Samsung Electro-Mechanics Co., Ltd. Light emitting diode package with diffuser and method of manufacturing the same
US7709857B2 (en) 2007-04-24 2010-05-04 Samsung Electro-Mechanics Co., Ltd. Light emitting diode package
US7759140B2 (en) 2003-10-21 2010-07-20 Samsung Led Co., Ltd. Light-emitting device and method of manufacturing the same
US7940350B2 (en) 2007-05-15 2011-05-10 Samsung Led Co., Ltd. Plane light source and LCD backlight unit having the same
US7959312B2 (en) 2006-12-05 2011-06-14 Samsung Led Co., Ltd. White light emitting device and white light source module using the same
US7964881B2 (en) 2007-10-19 2011-06-21 Samsung Led Co., Ltd. Semiconductor light emitting device, method of manufacturing the same, and semiconductor light emitting device package using the same
US7994525B2 (en) 2005-10-07 2011-08-09 Samsung Led Co., Ltd. Nitride-based semiconductor light emitting diode
US8008683B2 (en) 2008-10-22 2011-08-30 Samsung Led Co., Ltd. Semiconductor light emitting device
US8013352B2 (en) 2006-01-10 2011-09-06 Samsung Led Co., Ltd. Chip coated light emitting diode package and manufacturing method thereof
US8129711B2 (en) 2007-07-12 2012-03-06 Samsung Led Co., Ltd. Nitride semiconductor light emitting device and fabrication method thereof
US8179938B2 (en) 2008-09-16 2012-05-15 Samsung Electronics Co., Ltd. Light-emitting element capable of increasing amount of light emitted, light-emitting device including the same, and method of manufacturing light-emitting element and light-emitting device
US8386808B2 (en) 2008-12-22 2013-02-26 Intel Corporation Adaptive power budget allocation between multiple components in a computing system
US8399944B2 (en) 2002-09-02 2013-03-19 Samsung Electronics Co., Ltd. Light emitting diode and method for fabricating the same
US8433662B2 (en) 2009-05-15 2013-04-30 The Aerospace Corporation Systems and methods for a core management system for parallel processing of an evolutionary algorithm
US8459832B2 (en) 2008-09-10 2013-06-11 Samsung Electronics Co., Ltd. Light emitting device and system providing white light with various color temperatures
US8502242B2 (en) 2008-08-05 2013-08-06 Samsung Electronics Co., Ltd. Light emitting device, light emitting system having the same, and fabricating method of the light emitting device and the light emitting system
KR20130097505A (en) 2012-02-24 2013-09-03 삼성전자주식회사 Apparatus and method for controlling power of electronic device having multi-core
US8735931B2 (en) 2006-04-17 2014-05-27 Samsung Electronics Co., Ltd. Light emitting diode package and fabrication method thereof
US9069555B2 (en) 2011-03-21 2015-06-30 Intel Corporation Managing power consumption in a multi-core processor
US20150185816A1 (en) 2013-09-23 2015-07-02 Cornell University Multi-core computer processor based on a dynamic core-level power management for enhanced overall power efficiency
US20150241947A1 (en) 2014-02-25 2015-08-27 International Business Machines Corporation Distributed power management with performance and power boundaries
US20160062447A1 (en) 2014-08-27 2016-03-03 Mediatek Inc. Method and computer-readable medium for dynamically managing power of multi-core processor
US20160179162A1 (en) 2014-12-22 2016-06-23 Intel Corporation Holistic global performance and power management
US20160187944A1 (en) 2014-12-26 2016-06-30 Intel Corporation Dynamic hierarchical performance balancing of computational resources
US9880599B1 (en) 2016-10-12 2018-01-30 International Business Machines Corporation Priority-aware power capping for hierarchical power distribution networks

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7412353B2 (en) * 2005-09-28 2008-08-12 Intel Corporation Reliable computing with a many-core processor
US8010822B2 (en) * 2008-03-28 2011-08-30 Microsoft Corporation Power-aware thread scheduling and dynamic use of processors
CN101354664B (en) * 2008-08-19 2011-12-28 中兴通讯股份有限公司 Method and apparatus for interrupting load equilibrium of multi-core processor
WO2012069881A1 (en) * 2010-11-25 2012-05-31 Freescale Semiconductor, Inc. Method and apparatus for managing power in a multi-core processor

Patent Citations (52)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6645830B2 (en) 1996-08-27 2003-11-11 Seiko Epson Corporation Exfoliating method, transferring method of thin film device, and thin film device, thin film integrated circuit device and liquid crystal display device produced by the same
US6818530B2 (en) 1996-08-27 2004-11-16 Seiko Epson Corporation Exfoliating method, transferring method of thin film device, and thin film device, thin film integrated circuit device, and liquid crystal display device produced by the same
US6372608B1 (en) 1996-08-27 2002-04-16 Seiko Epson Corporation Separating method, method for transferring thin film device, thin film device, thin film integrated circuit device, and liquid crystal display device manufactured by using the transferring method
USRE38466E1 (en) 1996-11-12 2004-03-16 Seiko Epson Corporation Manufacturing method of active matrix substrate, active matrix substrate and liquid crystal display device
US7781727B2 (en) 1998-11-25 2010-08-24 Nuvotronics, Llc Optoelectronic component comprising an encapsulant
US7288758B2 (en) 1998-11-25 2007-10-30 Rohm And Haas Electronic Materials Llc Wafer-level optoelectronic device substrate
US7208725B2 (en) 1998-11-25 2007-04-24 Rohm And Haas Electronic Materials Llc Optoelectronic component with encapsulant
US8049161B2 (en) 1998-11-25 2011-11-01 Samsung Electronics Co., Ltd. Optoelectronic component with flip-chip mounted optoelectronic device
US6711691B1 (en) 1999-05-13 2004-03-23 Apple Computer, Inc. Power management for computer systems
US7087932B2 (en) 2000-07-18 2006-08-08 Sony Corporation Semiconductor light-emitting device and semiconductor light-emitting device
US6818465B2 (en) 2001-08-22 2004-11-16 Sony Corporation Nitride semiconductor element and production method for nitride semiconductor element
US6858081B2 (en) 2002-01-17 2005-02-22 Sony Corporation Selective growth method, and semiconductor light emitting device and fabrication method thereof
US6967353B2 (en) 2002-01-18 2005-11-22 Sony Corporation Semiconductor light emitting device and fabrication method thereof
US8399944B2 (en) 2002-09-02 2013-03-19 Samsung Electronics Co., Ltd. Light emitting diode and method for fabricating the same
US8536604B2 (en) 2002-09-02 2013-09-17 Samsung Electronics Co., Ltd. Light emitting diode and method for fabricating the same
US7002182B2 (en) 2002-09-06 2006-02-21 Sony Corporation Semiconductor light emitting device integral type semiconductor light emitting unit image display unit and illuminating unit
US7759140B2 (en) 2003-10-21 2010-07-20 Samsung Led Co., Ltd. Light-emitting device and method of manufacturing the same
US7319044B2 (en) 2003-12-23 2008-01-15 Samsung Electro-Mechanics Co., Ltd. Nitride semiconductor light emitting device and method for manufacturing the same
US7154124B2 (en) 2003-12-23 2006-12-26 Samsung Electro-Mechanics Co., Ltd. Nitride semiconductor light emitting device and method for manufacturing the same
US7281146B2 (en) 2004-06-30 2007-10-09 Intel Corporation Dynamic power requirement budget manager
US7084420B2 (en) 2004-10-26 2006-08-01 Samsung Electro-Mechanics Co., Ltd. Nitride based semiconductor device
US7501656B2 (en) 2005-07-26 2009-03-10 Samsung Electro-Mechanics Co., Ltd. Light emitting diode package with diffuser and method of manufacturing the same
US7790482B2 (en) 2005-07-26 2010-09-07 Samsung Led Co., Ltd. Light emitting diode package with diffuser and method of manufacturing the same
US7994525B2 (en) 2005-10-07 2011-08-09 Samsung Led Co., Ltd. Nitride-based semiconductor light emitting diode
US8324646B2 (en) 2006-01-10 2012-12-04 Samsung Electronics Co., Ltd. Chip coated light emitting diode package and manufacturing method thereof
US8013352B2 (en) 2006-01-10 2011-09-06 Samsung Led Co., Ltd. Chip coated light emitting diode package and manufacturing method thereof
US20070198863A1 (en) * 2006-02-17 2007-08-23 Pradip Bose Method and system for controlling power in a chip through a power-performance monitor and control unit
US8735931B2 (en) 2006-04-17 2014-05-27 Samsung Electronics Co., Ltd. Light emitting diode package and fabrication method thereof
US7959312B2 (en) 2006-12-05 2011-06-14 Samsung Led Co., Ltd. White light emitting device and white light source module using the same
US7709857B2 (en) 2007-04-24 2010-05-04 Samsung Electro-Mechanics Co., Ltd. Light emitting diode package
US8432511B2 (en) 2007-05-15 2013-04-30 Samsung Electronics Co., Ltd. Plane light source and LCD backlight unit having the same
US7940350B2 (en) 2007-05-15 2011-05-10 Samsung Led Co., Ltd. Plane light source and LCD backlight unit having the same
US8129711B2 (en) 2007-07-12 2012-03-06 Samsung Led Co., Ltd. Nitride semiconductor light emitting device and fabrication method thereof
US7985976B2 (en) 2007-10-19 2011-07-26 Samsung Led Co., Ltd. Semiconductor light emitting device, method of manufacturing the same, and semiconductor light emitting device package using the same
US8263987B2 (en) 2007-10-19 2012-09-11 Samsung Electronics Co., Ltd. Semiconductor light emitting device, method of manufacturing the same, and semiconductor light emitting device package using the same
US7964881B2 (en) 2007-10-19 2011-06-21 Samsung Led Co., Ltd. Semiconductor light emitting device, method of manufacturing the same, and semiconductor light emitting device package using the same
US8502242B2 (en) 2008-08-05 2013-08-06 Samsung Electronics Co., Ltd. Light emitting device, light emitting system having the same, and fabricating method of the light emitting device and the light emitting system
US8766295B2 (en) 2008-08-05 2014-07-01 Samsung Electronics Co., Ltd. Light emitting device, light emitting system having the same, and fabricating method of the light emitting device and the light emitting system
US8459832B2 (en) 2008-09-10 2013-06-11 Samsung Electronics Co., Ltd. Light emitting device and system providing white light with various color temperatures
US8179938B2 (en) 2008-09-16 2012-05-15 Samsung Electronics Co., Ltd. Light-emitting element capable of increasing amount of light emitted, light-emitting device including the same, and method of manufacturing light-emitting element and light-emitting device
US8008683B2 (en) 2008-10-22 2011-08-30 Samsung Led Co., Ltd. Semiconductor light emitting device
US8386808B2 (en) 2008-12-22 2013-02-26 Intel Corporation Adaptive power budget allocation between multiple components in a computing system
US8433662B2 (en) 2009-05-15 2013-04-30 The Aerospace Corporation Systems and methods for a core management system for parallel processing of an evolutionary algorithm
US9069555B2 (en) 2011-03-21 2015-06-30 Intel Corporation Managing power consumption in a multi-core processor
US9477296B2 (en) 2012-02-24 2016-10-25 Samsung Electronics Co., Ltd. Apparatus and method for controlling power of electronic device having multi-core
KR20130097505A (en) 2012-02-24 2013-09-03 삼성전자주식회사 Apparatus and method for controlling power of electronic device having multi-core
US20150185816A1 (en) 2013-09-23 2015-07-02 Cornell University Multi-core computer processor based on a dynamic core-level power management for enhanced overall power efficiency
US20150241947A1 (en) 2014-02-25 2015-08-27 International Business Machines Corporation Distributed power management with performance and power boundaries
US20160062447A1 (en) 2014-08-27 2016-03-03 Mediatek Inc. Method and computer-readable medium for dynamically managing power of multi-core processor
US20160179162A1 (en) 2014-12-22 2016-06-23 Intel Corporation Holistic global performance and power management
US20160187944A1 (en) 2014-12-26 2016-06-30 Intel Corporation Dynamic hierarchical performance balancing of computational resources
US9880599B1 (en) 2016-10-12 2018-01-30 International Business Machines Corporation Priority-aware power capping for hierarchical power distribution networks

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
TAPE: Thermal-Aware Agent-Based Power Economy for Multi/Many-Core Architectures,Thomas Ebi, Mohammad Abdullah Al Faruque, Joerg Henkel,ICCAD 2009.

Also Published As

Publication number Publication date
CN108509014A (en) 2018-09-07
KR20180098904A (en) 2018-09-05
CN108509014B (en) 2023-09-29
US10409351B2 (en) 2019-09-10
US20190339760A1 (en) 2019-11-07
US20180246554A1 (en) 2018-08-30

Similar Documents

Publication Publication Date Title
US11301016B2 (en) Computing devices and methods of allocating power to plurality of cores in each computing device
JP6097444B2 (en) System and method for memory system management based on temperature information of memory system
KR101748747B1 (en) Controlling configurable peak performance limits of a processor
JP6496626B2 (en) Heterogeneous integrated memory unit and its extended integrated memory space management method
US10255118B2 (en) Processing system including a plurality of cores and method of operating the same
Durand et al. Euroserver: Energy efficient node for european micro-servers
US20060294401A1 (en) Power management of multiple processors
US20160154734A1 (en) Cache memory device and electronic system including the same
US10768684B2 (en) Reducing power by vacating subsets of CPUs and memory
US8457805B2 (en) Power distribution considering cooling nodes
US8543770B2 (en) Assigning memory to on-chip coherence domains
US20200341673A1 (en) Intra-device notational data movement system
US11726834B2 (en) Performance-based workload/storage allocation system
US20230141508A1 (en) Methods and apparatus for auto tuning cooling of compute devices based on workloads
US11334390B2 (en) Hyper-converged infrastructure (HCI) resource reservation system
US9021281B2 (en) Run-time task-level dynamic energy management
US20230209772A1 (en) Methods and apparatus for resource lifetime aware cooling systems
US20220214965A1 (en) System and method for storage class memory tiering
US20230333612A1 (en) System and method for thermal management of multi-enclosure system
US20230195528A1 (en) Method and apparatus to perform workload management in a disaggregated computing system
US20230394364A1 (en) Reconfigurable computing fabric for machine learning processing
CN115729438A (en) Data access method, device and storage medium

Legal Events

Date Code Title Description
FEPP Fee payment procedure

Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY

STPP Information on status: patent application and granting procedure in general

Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION

STPP Information on status: patent application and granting procedure in general

Free format text: NON FINAL ACTION MAILED

STPP Information on status: patent application and granting procedure in general

Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER

STPP Information on status: patent application and granting procedure in general

Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS

STPP Information on status: patent application and granting procedure in general

Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT VERIFIED

STCF Information on status: patent grant

Free format text: PATENTED CASE